Systems and methods for application of statistical classification and pattern recognition for compartment design in horizontal oil wells

Embodiments described herein generally relate to systems and methods for application of statistical classification and pattern recognition for compartment design in horizontal oil wells and, more specifically, to embodiments for designing compartments in a horizontal oil well.

Horizontal oil wells provide increased reservoir contact and higher well productivity, but horizontal oil wells can suffer with uneven inflow along the well trajectory due to reservoir heterogeneity, resulting in premature water breakthrough. Therefore, horizontal oil wells completions often involve installing inflow control devices (ICD) for inflow equalization and delaying water breakthrough thereby increasing oil recovery from the wells.

The process of designing ICD completions involves placement of packers, such as reservoir isolation devices, at appropriate intervals of the horizontal section of the well. This establishes a number of inflow compartments in the completions design, as well as a length of each individual compartment. Appropriate design of compartments helps ensure effective inflow equalization along the well trajectory from target entry (TE) depth to target depth (TD). The higher the reservoir heterogeneity, the greater the number of compartments may be required for effective inflow equalization.

The traditional approach usually involves an engineer identifying packer placement locations by visual analysis of the well permeability log plotted on a log scale along the measured depth (MD) of the well trajectory. The engineer would normally create compartment intervals grouping similar magnitude of permeability.

This traditional approach has the challenge of being subjective to the interpretation of an engineer. Further, in a forecast optimization study utilizing reservoir simulation, multiple well completions designs may need to be developed with a custom design for each well due to the variation in reservoir permeability at each well location. As a field development plan may involve hundreds of wells, the process of designing compartments customized to each well can be a very tedious task.

Systems and methods for application of statistical classification and pattern recognition for compartment design in horizontal oil wells are described. One embodiment includes a drill for drilling a target well and a computing device that includes a memory component that stores logic that causes the computing device to receive an input parameter for the target well, perform a log transformation on the permeability log to create transformed data, and calculate a mean and a standard deviation of the transformed data. Some embodiments generate a classification flag that classifies the permeability log, based on the standard deviation, classify noise from the permeability log, and transform the noise based on a predefined pattern library. Some embodiments create a final transformed signal from the classification flag and the noise and generate a compartment design from the final transformed signal that provides recommended compartment intervals versus measured depth of the target well.

In another embodiment, a method includes receiving, by a computing device, an input parameter for a target well, where the input parameter includes a permeability log for the target well, performing, by the computing device, a log transformation on the permeability log to create transformed data, and calculating, by the computing device, a mean and a standard deviation of the transformed data. In some embodiments, the method includes generating, by the computing device, a classification flag that classifies the permeability log, based on the standard deviation, classifying, by the computing device, noise from the permeability log, and transforming, by the computing device, the noise based on a predefined pattern library. In some embodiments, the method includes creating, by the computing device, a final transformed signal from the classification flag and the noise and generating, by the computing device, a compartment design from the final transformed signal that provides a recommended compartment interval versus measured depth of the target well.

In yet another embodiment, a system includes drilling equipment for drilling a target well, wherein the target well is a horizontal well, where the drilling equipment includes a special purpose computing device that includes a memory component. The memory component may store logic, that when executed by the special purpose computing device, causes the system to receive an input parameter for the target well, where the input parameter includes a permeability log for the target well, perform a log transformation on the permeability log to create transformed data, and calculate a mean and a standard deviation of the transformed data. In some embodiments, the logic may cause the system to generate a classification flag that classifies the permeability log, based on the standard deviation, classify noise from the permeability log, and transform the noise based on a predefined pattern library. In some embodiments, the logic causes the system to create a final transformed signal from the classification flag and the noise and generate a compartment design from the final transformed signal that provides recommended compartment intervals versus measured depth of the target well.

These and additional features provided by the embodiments of the present disclosure will be more fully understood in view of the following detailed description, in conjunction with the drawings.

Embodiments disclosed herein include systems and methods for application of statistical classification and pattern recognition for compartment design in horizontal oil wells. Embodiments provided herein address challenges described above by a process of compartmentalization of a permeability log based on statistical analysis and signal processing techniques. These embodiments perform automatic transformation of the permeability data to recommend placement of packers along the measured depth of the well trajectory. This solves both the challenge of human subjectivity and tedious nature of designing multiple well completions through standardization and automation.

More specifically, horizontal oil wells with complex completions, specifically employing inflow control devices (ICD), provide the potential to improve oil recovery through inflow equalization and delayed water breakthrough. For effective inflow equalization, the horizontal section of the well may be divided into compartments based on variation of reservoir properties such as permeability.

Accordingly, some embodiments provide a computing device that may design well inflow compartments without human intervention. In some embodiments, the computing device performs a statistical analysis of a permeability log for the target well which follows in nature a log normal distribution. The computing device may perform a log transformation on the permeability data. Then, the computing device may calculate mean and standard deviation of transformed data which is used in calculation of a new property, log perm mean deviation, by subtracting the mean from the log permeability. The computing device may create a flag signal classifying the log permeability based on intervals of standard deviation. These embodiments may additionally transform a classification flag created at the log sampling resolution to the resolution of one joint length of a completion design using a sliding window process employing a most-of methodology.

The signal may be further transformed with the application of pattern recognition that identifies at least one interval below a user-specified minimum compartment length and automatically classifies the noise based on a pattern of interval−1 and interval+1 value and further transforms the noise based on a predefined pattern library.

Additionally, a compartment design array is generated from the transformed signal that forms the basis of automatic generation of compartment intervals along the measured depth of the target well. The systems and methods for application of statistical classification and pattern recognition for compartment design in horizontal oil wells incorporating the same will be described in more detail, below.

1 FIG. 100 102 104 Referring now to the drawings,depicts a computing environment for application of statistical classification and pattern recognition for compartment design in horizontal oil wells, according to embodiments provided herein. As illustrated, the drilling environment includes a networkthat is coupled to both a computing deviceand drilling equipment.

100 100 102 104 100 104 The networkmay be configured as any wide area network (such as PSTN, mobile network, the internet, satellite network, etc.), local network (such as wireless fidelity, local area network, etc.) and/or any peer to peer network (such as a ZigBee, near field communication, a wired connection, etc.). Regardless, the networkmay permit the computing deviceto be located remote from the drilling equipment; however some embodiments are configured with the networkrepresenting a connection of an integrated, special purpose computer with the drilling equipment.

102 102 140 144 144 102 102 104 a b 14 FIG. 1 FIG. The computing devicemay be configured as a personal computer, laptop, tablet, mobile device, integrated special purpose computer, programmable logic controller (PLC) and/or other type of computing device for providing the functionality provided herein. As such, the computing devicemay include a memory componentthat stores sensing logicand determining logic. Additional components of the computing deviceare described with reference to. It should be noted that while depicted inas a general purpose computer, some embodiments of the computing devicemay be a special purpose computing device that is integrated into the drilling equipment.

104 106 104 144 102 b The drilling equipmentmay be configured as any hardware for drilling a horizontal well into a reservoir, such as a hydrocarbon reservoir. The drilling equipmentmay include one or more hardware components for drilling the well and installing an ICD completion, as described herein. The determining logicmay cause the computing deviceto utilize that data to predict a likelihood of success for one or more logging tools, as described in more detail below.

1 FIG. It should be understood that while not explicitly depicted in, one or more sensors may be utilized for collecting input data. The sensor includes any of a plurality of different sensors for detecting temperature, pressure, salinity, length, depth, mud type, weight, viscosity, solids, electrical stability, etc.

2 FIG. 1 FIG. 104 106 202 202 204 204 206 208 208 210 a e a c a c depicts components of a horizontal well ICD completion, according to embodiments provided herein. As illustrated, the drilling equipmentfromdrilled the reservoirand installed infrastructure for the ICD completion. The infrastructure includes packers-, ICD devices-, a blank pipe compartment, flow compartments-, and an ICD joint.

208 202 204 Designing flow compartmentsmay be part of an ICD completions design for a horizontal well. Embodiments may identify intervals of similar well properties such as permeability along the horizontal section of the target well. The intervals may be designed to be isolated with each other through the installation of packers. Embodiments may additionally decide on a number of ICD devicesto be placed in each compartment and their respective opening choke sizes.

208 208 208 206 204 206 The total number of flow compartmentsand the size of each flow compartmentmay be related to heterogeneity of measured permeability property along a horizontal section of the well. For example, a heterogeneous well permeability may require a greater number of flow compartmentsof varying size, compared to a well with a lower heterogeneous permeability property. In addition, any intervals of permeability magnitude less than a predefined cut off value are considered blank pipe compartments, which means that no ICD devicesare placed in these blank pipe compartments.

3 FIG. 300 310 312 300 310 312 312 3120 a depicts a plotof a well permeability log versus measured depth, and a plotof a well permeability log versus measured depth, with the placement of packersgrouping similar permeability magnitudes into one interval of measured depth, according to embodiments provided herein. As illustrated in the plot, there are several depths of the well where the permeability drops to a very low value, relative to the permeability of the remaining depths of the well. Specifically referring to the plot, packersmay be placed at those regions (e.g,-where the permeability is low.

106 Specifically, as the fluid flow in a porous subsurface reservoir, given by Darcy's law, is directly proportional to the permeability property, approximately uniform fluid flow is expected from a section of wellbore with similar permeability magnitude. Hence, for effective inflow equalization along the wellbore, embodiments provided herein compartmentalize sections of the reservoirwith similar magnitude of permeability so that inflow control devices (ICD) can be deployed with customized inflow settings.

An actual permeability log from a real oilfield well is typically heterogeneous, which means that the magnitude of the reservoir property varies continuously along the measured depth of the wellbore. Therefore, statistical analysis of the permeability log is utilized to standardize the process of compartmentalization. The rock permeability property is generally considered to be log-normally distributed. Embodiments provided herein utilize this characteristic of permeability to analyze and classify the measured data along the wellbore.

4 4 FIGS.A,B 4 FIG.A 4 FIG.B 4 FIG.A 402 102 402 404 404 406 404 404 a b b c. depict compartment designs for two example wells, according to embodiments provided herein. Specifically, the compartment design ofrepresents a less heterogeneous well that requires a smaller number of compartments. The compartment design ofrepresents a more heterogeneous well that requires more compartments for effective ICD design. As illustrated in, the permeability versus measured depth is depicted in plot. Embodiments of the computing devicemay be configured to analyze the plot(and/or related data) and design the completions with a first flow compartmentdefining the depth between two low areas of permeability. The embodiments may define a second flow compartmentbetween another two areas of low permeability. A bland pipe compartmentmay be defined between the second flow compartmentand a third flow compartment

4 FIG.B 412 414 416 414 416 414 414 416 414 414 a a b b b c c c d. As illustrated in, a plotof a permeability versus measured depth for a second well is provided. In this embodiment, the second well is more heterogeneous, thus dictating a different completion design. Specifically, embodiments may define a first flow compartment, which is adjacent to a first blank pipe compartment. A second flow compartmentmay then be defined between lower permeability sections of the well. A second blank pipe compartmentmay then be disposed between the second flow compartmentand a third flow compartment. A third blank pipe compartmentmay be adjacent to the third flow compartmentand a fourth flow compartment

102 Embodiments are provided herein to smartly define flow compartments automatically without human intervention. To this end, the computing devicetakes the following parameters as input: well permeability log file (permeability versus measured depth), well target entry depth (TE-DEPTH), length of a completion tubing joint (J), minimum length of a compartment (L), permeability cut-off value (K-MIN), and/or other data.

In these embodiments, the design of flow compartments in a parameterized process, such that multiple realizations of compartments design are swiftly generated. For example, varying the value of parameter L results in different unique compartment designs, which as part of the overall ICD design process results in different unique ICD completions designs.

In some embodiments, this functionality may be deployed as part of reservoir simulation based automated ICD design optimization workflow (such as the one described in US Patent, US 2021/0350035, which is incorporated by reference in its entirety).

5 FIG.A 5 FIG.B 5 FIG.C 502 502 10 506 depicts a histogramof well permeability log data plotted on natural scale, according to embodiments provided herein. As illustrated, the histogramof a typical permeability log is shown in natural scale on the left where it does not exhibit a normal distribution. However, when the data is log transformed to the basein, the distribution appears to be a normal distribution. An estimated probability density function (PDF) plotted on the log transformed histograminconfirms a slightly left skewed log-normal distribution.

5 FIG.B 504 depicts a histogramof log-transformed data that demonstrates a normal distribution, according to embodiments provided herein. The general form of probability density function of a normal distribution function is given as:

6 6 FIGS.A,B where f(x) is the probability density function, a is the standard deviation and μ is the mean., discussed below, shows the characteristics of a normal distribution described with mean and standard deviation.

5 FIG.C 5 FIG.B 504 504 depicts an estimated probability density function plotted on the log-transformed histogram, according to embodiments provided herein. As illustrated above, Equation 1 may be utilized to determine the estimated PDF of the histogramfrom.

6 FIG.A 602 604 606 608 610 612 a a a a a − a depicts a normal distribution showing mean and intervals of standard deviation from a mean, with the percentage numbers representing the empirical rule of a normal distribution, according to embodiments provided herein. In order to efficiently compartmentalize the wellbore for complex completions, embodiments provided herein perform standard-deviation (SD) based classification. The permeability valuesthat lie above the mean and up to 1 SD are marked as Class 1, the valuesthat are one SD above the mean are marked as Class 2, the valuesthat are two SD above the mean are marked as Class 3. The valuesbelow the mean and up to −1 SD are marked as Class −1. The values1 SD below the mean and up to −2SD are marked as Class −2. All valuesbelow −2S D are marked as Class −3. The classification method effectively identifies groups of permeability with similar magnitude. The class values from −3 to 2 are known as Class IDs.

6 FIG.B 602 604 608 610 612 b b b b − b depicts a standard deviation curve, with positive class divisions representing data higher than the mean and negative class divisions representing values lower than the mean, according to embodiments provided herein. As illustrated, in order to efficiently compartmentalize the wellbore for complex completions, the methodology in this invention performs standard-deviation (SD) based classification. The permeability valuesthat lie above the mean and up to 1 SD are marked as Class 1. The values1 SD above the mean are marked as Class 2. The valuesbelow the mean and up to −1 SD are marked as Class −1. The values1 SD below the mean and up to −2 SD are marked as Class −2. All valuesbelow −2SD are marked as Class −3. The classification method hence effectively identifies groups of permeability with similar magnitude. The class values from −3 to 2 are known as Class IDs. This concept is applied to the automated compartment design process described below.

7 FIG. 700 710 720 106 102 depicts a plotof distribution of oilfield well permeability versus measured depth, a plotof log-transformed distribution of oilfield well permeability versus measured depth, and a plotof mean deviation distribution of oilfield well permeability versus measured depth, according to embodiments provided herein. As illustrated, there are typically two types of compartment intervals in an ICD completions design of a horizontal well. Specifically, compartments that have ICD for inflow equalization and blank pipe compartments where no ICD devices are installed. The later compartment type is usually designed for intervals where permeability is very low below the permeability cut off value for the reservoir. The computing devicemay be configured to automatically identify both types of compartment intervals along the horizontal section of an oilfield wellbore.

700 10 710 x−μ At least one embodiment takes permeability data with respect to the MD of the wellbore (log) as input (plot) and performs logarithmic transformation to the data with a logarithm of base(plot). Any zero or undefined values are replaced with the minimum non-zero value of the data before log transformation. Then, the mean and standard deviation of the data are calculated and then, the mean deviation is calculated, known as LOGMEANDEV, for every point in the dataset as below;LOGMEANDEV=  Equation (2)

where x is the log-transformed permeability data.

8 FIG. 7 FIG. 7 FIG. 800 810 820 830 800 710 102 depicts a plotfor class flag versus measured depth plot, a plotfor class flag 1JL versus measured depth plot, a plotfor unique intervals versus measured depth plot, and a plotfor normalized individual lengths versus measured depth plot, according to embodiments provided herein. Continuing from the process in, these embodiments create a classifier array known as the CLASS_FLAG (plot), where for each value of LOGMEANDEV a sample is classified based on standard deviation classification from plot(). The computing devicealso receives an input of a permeability cut off value from user and any values below the cut off are assigned a Class −4, which identifies blank pipe compartment intervals. A data point before or after the data array (null interval) is considered a class ID value of 0.

Embodiments may additionally calculate maximum compartment intervals, “Imax”, that can be physically installed in the target wellbore which is a function of the joint length of completions tubing taken as an input from the user. The relationship to calculate is as follows:

where “J” is the joint length given in feet, “S” is the number of samples per feet and “N” is the total number of data sample values in the input permeability array.

102 800 102 810 The computing devicemay additionally smooth the CLASS_FLAG array (plot) using sliding window process employing a “most-of” methodology. The selected window size may be equal to about 1 joint length. For each window, the computing devicecalculates a maximum number of samples of each Class ID and assigns a new value to the window equal to the Class ID with maximum samples. The new array is known as CLASS_FLAG_1JL (plot).

810 102 830 820 102 830 820 The CLASS_FLAG_1JL array (plot) identifies intervals of similar permeability magnitude (SD-based Classes) with a length equal to about one joint length of tubing. However, this may not be suitable to be used for compartments design as it can result in a recommendation for a very large number of compartments especially for heterogeneous wellbores. In some operational deployments, a large number of compartments are not desired due to operational constraints and cost. Therefore, embodiments further process the CLASS_FLAG_1JL to reduce the number of compartments to practical level. To process this array further, the computing devicecalculates a plurality of new arrays, or the number of unique intervals in the CLASS_FLAG_1JL array, UNIQ_INT (plot) and the corresponding length of each unique interval, INT_LEN (plot). The computing devicealso receives a minimum compartment length as input from the user. Any UNIQ_INT (plot) that has a INT_LEN (plot) less than that the minimum compartment length is marked as noise that needs to be further processed.

9 9 FIGS.A-C 9 9 FIGS.A andB 902 depict plotsshowing signal patterns, according to embodiments provided herein. As illustrated, the x-axis represents the class IDs and “i−1,” “i+1” on the y-axis represent the intervals preceding and following the noise interval respectively. “o” represents the class value that is assigned to the noise based on the pattern of i−1, i+1. Please note that a class ID value of 0 represents a null interval. A null interval may include any point before or after data array. All 48 patterns in the library are shown in.

102 At this stage, the computing deviceemploys a pattern recognition technique to process noise (small intervals less than minimum compartment length) in CLASS_FLAG_1JL based on a pre-defined library of patterns. A pattern is comprised of Class ID values of the preceding interval “i−1” and the following interval “i+1”. Based on this pattern, a Class ID value is assigned to the noise (small interval) given by “o”. After a pattern is read, it is matched with pre-defined patterns in the pattern library and the noise interval is updated with the corresponding value of Class ID “o” associated with the pattern. As illustrated, “i−1” represents the Class ID of the preceding interval, “i+1” represents the Class ID of the following interval and “o” in red color represents the value that will be assigned to the noise (small interval) after pattern recognition processing.

10 FIG. 1000 1010 1020 1010 102 1020 1020 102 depicts a plot () of input permeability versus measured depth, a plot () of the final class flag versus measured depth after pattern recognition processing, and a plot () of an output compartment identifier array versus measured depth, according to embodiments provided herein. As illustrated, the pattern recognition processing results in a new final array known as CLASS_FLAG_PAT_R (plot) that groups similar magnitudes of permeability based on standard-deviation based classification. The computing devicemay additionally generate an output array known as COMP_NUM_AUTO (plot). The COMP_NUM_AUTO (plot) is exported as a 2-D array along with measured depth of the wellbore. The values of 1 to N specify the compartment number where N is the maximum compartments identified by the computing deviceand the value of 0 identifies blank pipe compartment intervals.

11 FIG. 1020 The output compartment identifier array can be utilized by users for manual design of ICD completions or can be used as an input to an automated ICD completions design software.illustrates the COMP_NUM_AUTO (plot) array being used in the ICD completions design for a horizontal well.

11 FIG. 10 FIG. 1100 1000 1020 102 1102 1100 1104 1100 1106 depicts a diagramof an inflow control device completions design versus measured depth, with compared with the plotfor input permeability and the plotfor final class flag from, according to embodiments provided herein. The computing devicemay at least one define reservoir isolation packer, which are represented in diagramas horizontal lines. Inflow completion devicesare represented in diagramwith round disc icons. Also depicted are nozzle ICDs (NICDs).

12 12 FIGS.A-C 7 8 10 FIGS.,, 12 12 FIGS.A-C 11 depict various plots for analysis provided in, and, applied to different wells, according to embodiments provided herein. As illustrated all three wells ofhave been processed with the process described above, using the following input parameters while the only change is the input permeability versus measured depth data.

Length of a completion tubing joint=38 ft.

Minimum length of a compartment=4.

Permeability cut-off value for blank pipe compartment=1 mD.

13 FIG. 1350 102 1352 1354 1356 depicts a flowchart for selecting a design simulation, according to embodiments provided herein. As illustrated in block, a reservoir simulation model is generated by the computing device. In block, using the simulation model, properties of a reservoir associated with a target well may be predicted, where the properties relate to a trajectory of the well. In block, a plurality of open hole design simulations are generated. In block, representations of a plurality of compartments in the target well may be generated, where the representations are automatically generated without user intervention using a synthetic production logging profile and the properties of the reservoir associated with the target well, where the synthetic production logging profile is based on fluid flow data estimations along the trajectory of the target well.

1358 1360 1362 1364 In block, a first design simulation and a second design simulation are generated automatically and without user intervention. The simulations may be based on the synthetic production logging profile and the properties of the profile of the target well. The first design simulation includes representations of a first plurality of n-flow devices and the second simulation includes different representations of a second plurality of n-flow devices may be based on density parameters and cross-sectional area parameters. In block, case selection factors may be generated automatically and without user intervention. The case selections factors may be associated with each of the first design simulation and the second design simulation. In block, the case selection factors may be ranked automatically and without user intervention. In block, at least one of the first design simulation and the second design simulation may be selected, based on the ranking, automatically and without user intervention.

14 FIG. 1450 depicts a flowchart for statistical classification and pattern recognition for compartment design in horizontal oil wells, according to embodiments provided herein. As illustrated in block, an input file may be read. As discussed above, the input file may include well permeability log file (permeability versus measured depth), well target entry depth (TE-DEPTH), length of a completion tubing joint (J), minimum length of a compartment (L), permeability cut-off value (K-MIN), and/or other data.

1452 1454 1456 1458 1460 1462 6 6 FIGS.A andB In block, the input parameters are read. In block, a standard deviation classification may be performed, as described with reference to. In block, a sliding window process may be executed to smooth the signal. In block, unique intervals and corresponding lengths of the ICD may be calculated. In block, a determination is made regarding whether the interval is less than a minimum interval length. The minimum length of a compartment may be measured as units of joint length. If not, at, no change in classification is made.

1460 1464 102 1466 9 9 FIGS.A andB If the interval is less than a minimum interval length at block, at block, the computing devicereads the patterns from. In block, the classification is updated with pattern recognition processing.

1468 1462 1466 1470 1472 102 In block, the information from blockand blockare merged. In block, a final processed classification is generated. In block, an output compartment recommendation file is written by the computing device.

15 FIG. 1550 1552 1554 depicts an additional flowchart for statistical classification and pattern recognition for compartment design in horizontal oil wells, according to embodiments provided herein. As illustrated in block, an input parameter for a target well may be received (such as user input), where the input parameter includes a well permeability log for the target well. In block, a log transformation is performed on the permeability log to create transformed data. In block, a mean and a standard deviation of the transformed data is performed.

1556 1558 1560 1562 1564 In block, a classification flag is generated that classifies the permeability log, based on the standard deviation. In block, noise from the permeability log may be classified. In block, the noise may be transformed, based on a predefined pattern library. In block, a final transformed signal may be created from the classification flag and the noise. In block, a compartment design array may be generated from the final transformed signal that provides recommended compartment intervals versus the measured depth of the target well.

16 FIG. 102 1430 1432 1434 1436 1438 1438 140 140 102 102 a b depicts a computing device for statistical classification and pattern recognition for compartment design in horizontal oil wells, according to embodiments provided herein. As illustrated, the computing deviceincludes a processor, input/output hardware, a network interface hardware, a data storage component(which stores well data, calculation data, and/or other data), and a memory component. The memory componentmay be configured as volatile and/or nonvolatile memory and as such, may include random access memory (including SRAM, DRAM, and/or other types of RAM), flash memory, secure digital memory, registers, compact discs (CD), digital versatile discs (DVD) (whether local or cloud-based), and/or other types of non-transitory computer-readable mediums. Depending on the particular embodiment, these non-transitory computer-readable mediums may reside within the computing deviceand/or external to the computing device.

140 1442 144 144 1446 102 a b 14 FIG. The memory componentmay store operating logic, the sensing logicand the determining logic. Each of these logic components may include a plurality of different pieces of logic, each of which may be embodied as a computer program, firmware, and/or hardware, as an example. A local interfaceis also included inand may be implemented as a bus or other communication interface to facilitate communication among the components of the computing device.

1430 1436 140 1432 The processormay include any processing component operable to receive and execute instructions (such as from a data storage componentand/or the memory component). As described above, the input/output hardwaremay include and/or be configured to interface with speakers, microphones, and/or other input/output components.

1434 102 The network interface hardwaremay include and/or be configured for communicating with any wired or wireless networking hardware, including an antenna, a modem, a LAN port, wireless fidelity (Wi-Fi) card, WiMAX card, mobile communications hardware, and/or other hardware for communicating with other networks and/or devices. From this connection, communication may be facilitated between the computing deviceand other computing devices.

1442 102 144 140 1430 108 144 1430 144 a b a The operating logicmay include an operating system and/or other software for managing components of the computing device. As discussed above, the sensing logicmay reside in the memory componentand may be configured to cause the processorto interpret signals from the sensor. The determining logicmay be configured to cause the processorto utilize the data from the sensing logicto likelihood of success for a desired logging option (as well as provide other functionality provided herein).

14 FIG. 102 102 102 144 144 a b It should be understood that while the components inare illustrated as residing within the computing device, this is merely an example. In some embodiments, one or more of the components may reside external to the computing deviceor within other devices. It should also be understood that, while the computing deviceis illustrated as a single device, this is also merely an example. In some embodiments, the sensing logicand the determining logicmay reside on different devices.

102 144 144 144 144 a b a b Additionally, while the computing deviceis illustrated with the sensing logicand the determining logicas separate logical components, this is also an example. In some embodiments, a single piece of logic may provide the described functionality. It should also be understood that while the sensing logicand the determining logicare described herein as the logical components, this is also an example. Other components may also be included, depending on the embodiment.

Accordingly, embodiments of this disclosure include the following aspects:

A first aspect includes a method for application of statistical classification and pattern recognition for compartment design in horizontal oil wells comprising: receiving, by a computing device, an input parameter for a target well, wherein the input parameter includes a permeability log for the target well; performing, by the computing device, a log transformation on the permeability log to create transformed data; calculating, by the computing device, a mean and a standard deviation of the transformed data; generating, by the computing device, a classification flag that classifies the permeability log, based on the standard deviation; classifying, by the computing device, noise from the permeability log; transforming, by the computing device, the noise based on a predefined pattern library; creating, by the computing device, a final transformed signal from the classification flag and the noise; and generating, by the computing device, a compartment design from the final transformed signal that provides a recommended compartment interval versus measured depth of the target well.

A second aspect includes the method the first aspect, wherein the input parameter further includes at least one of the following: a well target entry depth, a length of a completion tubing joint, a minimum length of a compartment, or a permeability cut-off value.

A third aspect includes the method of any of the first aspect and/or the second aspect, wherein the classification flag is created at a log sampling resolution.

A fourth aspect includes the method of any of the first aspect through the third aspect, further comprising transforming the classification flag to a resolution of one joint length of a completion design.

A fifth aspect includes the method of any of the first aspect through the fourth aspect, wherein performing the log transformation includes performing a statistical analysis of permeability log for the target well.

A sixth aspect includes the method of any of the first aspect through the fifth aspect, wherein the noise is classified based on a pattern of interval−1 and interval+1 value.

A seventh aspect includes the method of any of the first aspect through the sixth aspect, further comprising receiving user input of a permeability cut off value and values below a predetermined cutoff are assigned a class −4, which identifies blank pipe compartment intervals.

An eighth aspect includes the method of any of the first aspect through the seventh aspect, further comprising calculating a maximum compartment interval, that can be physically installed in the target well, wherein the maximum compartment interval is a function of a joint length of completions tubing received as user input.

max A ninth aspect includes the method of any of the first aspect through the eighth aspect, further comprising calculating the maximum compartment interval, Ias:

where “J” is a joint length given in feet, “S” is a number of samples per feet and “N” is a total number of data sample values in an input permeability array.

A tenth aspect includes the method of any of the first aspect through the ninth aspect, further comprising smoothing the classification flag using a sliding window with a selected window size equal to about 1 joint length.

An eleventh aspect includes the method of any of the first aspect through the tenth aspect, further comprising grouping similar magnitudes of permeability based on standard-deviation based classification to create the final transformed signal.

A twelfth aspect includes the method of any of the first aspect through the eleventh aspect, further comprising defining a reservoir isolation packer, an inflow completion device, and a nozzle inflow control device (NICD).

A thirteenth aspect includes a system for application of statistical classification and pattern recognition for compartment design in horizontal oil wells comprising: drilling equipment for drilling a target well, wherein the target well is a horizontal well, wherein the drilling equipment includes a special purpose computing device that includes a memory component that stores logic, that when executed by the special purpose computing device, causes the system to perform at least the following: receive an input parameter for the target well, wherein the input parameter includes a permeability log for the target well; perform a log transformation on the permeability log to create transformed data; calculate a mean and a standard deviation of the transformed data; generate a classification flag that classifies the permeability log, based on the standard deviation; classify noise from the permeability log; transform the noise based on a predefined pattern library; create a final transformed signal from the classification flag and the noise; and generate a compartment design from the final transformed signal that provides recommended compartment intervals versus measured depth of the target well.

A fourteenth aspect includes the system of the thirteenth aspect, wherein the input parameter further includes at least one of the following: a well target entry depth, a length of a completion tubing joint, a minimum length of a compartment, or a permeability cut-off value.

A fifteenth aspect includes system of any of the thirteenth and/or fourteenth aspect, wherein the noise is classified based on a pattern of interval−1 and interval+1 value.

A sixteenth aspect includes the system of any of the thirteenth aspect through the fifteenth aspect, wherein the logic further causes the system to receive user input of a permeability cut off value and values below a predetermined cutoff are assigned a class −4, which identifies blank pipe compartment intervals.

A seventeenth aspect includes the system of any of the thirteenth aspect through the sixteenth aspect, wherein the logic further causes the system to calculate a maximum compartment interval, that can be physically installed in the target well, wherein the maximum compartment interval is a function of a joint length of completions tubing received as user input.

max An eighteenth aspect includes the system of any of the thirteenth aspect through the seventeenth aspect, wherein the logic further causes the system to calculate the maximum compartment interval, Ias:

where “J” is a joint length given in feet, “S” is a number of samples per feet and “N” is a total number of data sample values in an input permeability array.

A nineteenth aspect includes drilling equipment for application of statistical classification and pattern recognition for compartment design in horizontal oil wells comprising: a drill for drilling a target well, wherein the target well is a horizontal well; and a computing device that includes a memory component that stores logic, that when executed by the computing device, causes the computing device to perform at least the following: receive an input parameter for the target well, wherein the input parameter includes a permeability log for the target well; perform a log transformation on the permeability log to create transformed data; calculate a mean and a standard deviation of the transformed data; generate a classification flag that classifies the permeability log, based on the standard deviation; classify noise from the permeability log; transform the noise based on a predefined pattern library; create a final transformed signal from the classification flag and the noise; and generate a compartment design from the final transformed signal that provides recommended compartment intervals versus measured depth of the target well.

max A twentieth aspect includes the drilling equipment of the nineteenth aspect, wherein the logic further causes the computing device to calculate a maximum compartment interval, that can be physically installed in the target well, wherein the maximum compartment interval is a function of a joint length of completions tubing received as user input, wherein the maximum compartment interval, Iis calculated as:

where “J” is a joint length given in feet, “S” is a number of samples per feet and “N” is a total number of data sample values in an input permeability array.

As illustrated above, various embodiments for application of statistical classification and pattern recognition for compartment design in horizontal oil wells are disclosed. Some embodiments include automatically designing well inflow compartments without any human intervention. These embodiments may be configured to perform a statistical analysis of the interpreted permeability log for the target well which is assumed to be log normally distributed. Some embodiments may perform a log transformation on the permeability data. Mean and standard deviation of transformed data may be calculated, which is used in calculation of a new property, log perm mean deviation, by subtracting the mean from the log permeability. A flag signal classifying the log permeability may be generated, based on intervals of standard deviation.

These embodiments may additionally transform the classification flag created at the log sampling resolution to the resolution of one joint length of a completion design using a sliding window process employing a most-of methodology. The signal may be transformed with the application of a pattern recognition methodology that identifies intervals below the user-specified minimum compartment length, then automatically classifies the noise based on a pattern of interval−1 and interval+1 value and in the end further transforms the noise based on a pre-defined pattern library. A compartment design array may also be generated from the final transformed signal that provides recommended compartment intervals versus the measured depth of the target well.

These embodiments may be configured to drastically increase the number of computations that can be performed for a single well or multiple wells. With at least 5-10 input parameters, as well as the other data computed by the computing device, the amount of information would be impossible for a human to calculate, especially considering that the drilling apparatus is highly expensive, so even minimal improvements in time save vast amounts of money.

While particular embodiments and aspects of the present disclosure have been illustrated and described herein, various other changes and modifications can be made without departing from the spirit and scope of the disclosure. Moreover, although various aspects have been described herein, such aspects need not be utilized in combination. Accordingly, it is therefore intended that the appended claims cover all such changes and modifications that are within the scope of the embodiments shown and described herein.

It should now be understood that embodiments disclosed herein include systems, methods, and non-transitory computer-readable mediums for application of statistical classification and pattern recognition for compartment design in horizontal oil wells. It should also be understood that these embodiments are merely exemplary and are not intended to limit the scope of this disclosure.